In recent years, the rapid growth of the Internet and mobile devices has led to a significant expansion of e-commerce, which has become deeply integrated into social life and made substantial contributions to the economy. To enhance the efficiency and diversity of e-commerce activities, there has been considerable research interest in the application of supportive techniques and technologies. This paper explores the use of machine learning to improve recommendation systems by analyzing purchase history and user emotions expressed in product reviews. A model was proposed and validated using datasets collected from e-commerce platforms. The study evaluated the model's effectiveness using Accuracy, Recall, and F1 metrics. Experimental results indicate that the SVM algorithm provides the best prediction performance, effectively aligning with user needs. These findings suggest that the model can be applied across various e-commerce platforms to enhance product and service recommendations for users.